The effect of human rhythm on packet delivery

In communication networks such as the Internet, the relationship between packet generation rate and time is similar to a rectangle wavefunction due to the rhythm of humans. Thus, we investigate the traffic dynamics on a network with a rectangle wavepacket generation rate. It is found that the critical delivering capacity parameter beta(c) (which separates the congested phase and the free phase) decreases significantly with the duty cycle r of the rectangle wave for package generation. And, in the congested phase, more collective generation of packets (smaller r) is helpful for decreasing the packet aggregation rate. Moreover, it is found that the congested phase can be divided into two regions, i.e., region1 and region2, where the distributions of queue lengths are nonlinear and linear, respectively. Also, the linear expression for the distribution of queue lengths in region2 is obtained analytically. Our work reveals an obvious effect of the rectangle wave on the traffic dynamics and the queue length distribution in the system, which is of essential interest and may provide insights into the designing of work-rest schedules and routing strategies.

基于Internet工作流的复合Web服务框架

实现灵活方便的企业业务集成一直是信息领域的核心问题，也是B2B电子商务应用的关键。为此将Web服务和传统的工作流技术相结合，设计并实现了支持复合Web服务运行和管理的框架WSFlow。给出了WSFlow的总体结构，描述了其中的关键技术，包括Web服务与工作流活动的动态配置和绑定技术，复合Web服务流程的动态修改以及复合Web服务的运行监控等技术。

基于支持向量机的Internet流量分类研究

准确的网络流量分类是众多网络研究工作的基础,也一直是网络测量领域的研究热点.近年来,利用机器学习方法处理流量分类问题成为了该领域一个新兴的研究方向.在目前研究中应用较多的是朴素贝叶斯(nave Bayes,NB)及其改进算法.这些方法具有实现简单、分类高效的特点.但该方法过分依赖于样本空间的分布,具有内在的不稳定性.因此,提出一种基于支持向量机(support vector machine,SVM)的流量分类方法.该方法利用非线性变换和结构风险最小化(structural risk minimization,SRM)原则将流量分类问题转化为二次寻优问题,具有良好的分类准确率和稳定性.在理论分析的基础上,通过在实际网络流集合上与朴素贝叶斯算法的对比实验,可以看出使用支持向量机方法处理流量分类问题,具有以下3个优势:1)网络流属性不必满足条件独立假设,无须进行属性过滤;2)能够在先验知识相对不足的情况下,仍保持较高的分类准确率;3)不依赖于样本空间的分布,具有较好的分类稳定性.

achieving on-time delivery: a two-stage probabilistic scheduling strategy for software projects

Chinese Academy of Sciences (ISCAS)